ABSTRACT
Panel data allows us to measure changes over time for countries, firms or individuals. In the last two chapters, we stressed that the advantage of panel data is that we can control for factors that are time-invariant and correlated with the variable of interest. In the case of firms, for example, we may think that firms with better managers use more inputs – so we cannot identify the effect of changing inputs. We also stressed that if there are time-varying unobservables, even fixed effect estimates will be biased.
